How Gas Fees Work and How to Reduce Them

What Gas Fees Are: Network Resources, Validators, and Transaction Priority

In 2024, Ethereum users paid about $3.7 billion in transaction fees, down from roughly $9.6 billion in 2023, reflecting lower average fees and shifting activity across scaling networks (data from Token Terminal). That change highlights what gas fees represent: a market price for scarce network resources, not a fixed “service charge”. When demand for block space rises, users compete by offering higher fees to secure faster inclusion.

Gas is the unit that measures computational work on networks such as Ethereum. A simple ETH transfer typically uses 21,000 gas, while a decentralised exchange swap often consumes 120,000–200,000 gas, depending on the contract path and token behaviour. Users pay gas used × gas price, so complex smart-contract calls cost more even when the gas price stays constant.

Validators (or miners on proof-of-work chains) prioritise transactions that pay more per unit of gas because block space remains limited. On Ethereum after EIP-1559, each transaction includes (1) a base fee that the protocol burns and (2) a priority fee (tip) that goes to the validator. The base fee adjusts automatically: when blocks exceed the target utilisation, it increases; when demand falls, it decreases. Ethereum targets about 15 million gas per block, with an elastic ceiling of roughly 30 million gas, which helps absorb short spikes but does not remove competition for inclusion.

• Network resources: computation, storage reads/writes, and calldata all consume gas, so contract design and transaction type drive cost.
• Validators: validators select transactions that maximise fee revenue while respecting block gas limits.
• Transaction priority: higher priority fees usually reduce confirmation time during congestion, while low tips risk delays or replacement.

How Gas Is Calculated: Gas Limit, Base Fee, and Priority Fee Mechanics

At 09:15 on a Tuesday, a user swaps 0.5 ETH on Uniswap and sees an estimated fee of 0.0019 ETH. The wallet shows three inputs: a gas limit of 150,000 units, a base fee of 20 gwei, and a priority fee of 2 gwei. The calculation follows a simple rule: fee paid equals gas used multiplied by the total gas price, so 150,000 × (20 + 2) gwei equals 3,300,000 gwei, or 0.0033 ETH at the maximum. If the swap consumes only 90,000 gas, the network charges 90,000 × 22 gwei, and the unused gas limit is not spent.

The base fee changes block by block under EIP-1559 and rises when blocks fill above the target utilisation. Ethereum targets 15 million gas per block and allows up to 30 million, so sustained demand pushes the base fee upward until activity slows (see EIP-1559). Validators do not receive the base fee; the protocol burns it, while the priority fee goes to validators as an incentive to include the transaction quickly.

Across transfers, swaps, and NFT mints, the same mechanics apply: set a realistic gas limit to avoid failed execution, then choose a priority fee that matches urgency. When the base fee spikes, waiting for a quieter block often reduces cost more than changing the tip.

Why Gas Fees Spike: Congestion, Block Space, and Market Dynamics

Gas fees spike when demand for block space rises faster than supply. Two common situations create that imbalance: a network-wide surge (many users compete for the same limited capacity) or a user-level urgency spike (a subset of users bids aggressively to get included sooner). Both push the market-clearing price higher, but they do so through different dynamics.

Option A: Congestion-driven spikes occur when activity jumps across the network at once. Examples include popular NFT mints, airdrop claims, or volatile price moves that trigger liquidations and arbitrage. Ethereum’s block capacity stays relatively stable from minute to minute, so a sudden increase in pending transactions forces users to outbid one another. Under EIP-1559, the protocol raises the base fee when blocks run above the target utilisation, which can lift fees for every user until demand cools.

Option B: Market-driven spikes happen when a smaller group needs fast inclusion and pays for priority. Searchers and bots often compete in “priority fee” auctions to capture time-sensitive opportunities, such as decentralised exchange price differences. That competition can raise the effective fee paid by users who set high tips, even if general network activity looks moderate.

Driver	What changes	Typical trigger	What users observe
Network congestion	Base fee rises as blocks fill	NFT mint, airdrop, market volatility	Wallet estimates climb for most transactions
Block space scarcity	Inclusion becomes competitive	Backlog of pending transactions	Longer confirmation times at “normal” fees
Priority bidding	Tips increase for faster inclusion	Arbitrage, liquidations, MEV competition	Higher “fast” option, wider fee range

Practical implications follow directly. When congestion drives the spike, waiting 10–30 minutes often reduces cost because the base fee falls as blocks return to target utilisation. When priority bidding dominates, paying a higher tip may not guarantee savings in time-sensitive periods, so users benefit more from avoiding peak events and using fee controls in wallets such as MetaMask to cap maximum fees.

How to Estimate Gas Before Sending: Wallet Tools, Explorers, and Fee Models

Underpaying gas can leave a transaction pending for hours, while overpaying can add several pounds to a simple transfer. As of February 2026, a standard Ethereum transfer typically uses 21,000 gas, so a 10 gwei pricing error changes the cost by 0.00021 ETH per transfer. At an ETH price of £2,000, that is about £0.42 wasted or underbid each time.

Estimate gas by combining wallet previews, explorer data, and the EIP-1559 fee model. Most wallets show a live fee quote and an expected confirmation time; treat that quote as a starting point, not a guarantee. Cross-check the current base fee and typical priority fee on an explorer such as Etherscan, then sanity-check against a public tracker such as ethereum.org guidance on gas mechanics.

Implement a three-step check before sending. Set the gas limit from the wallet’s simulation (for example, 21,000 for a transfer, 120,000–200,000 for many swaps). Next, set a max fee per gas that covers short-term volatility; a common approach is base fee plus 2–3 gwei for the tip, plus a buffer of 20–30% on the max fee. Then, if the transaction remains pending beyond 2–3 blocks, replace it with a higher max fee using the wallet’s “speed up” function.

This workflow reduces failed or stuck transactions and caps overspend. In practice, keeping the tip within 1–3 gwei during normal conditions often achieves inclusion within a few blocks without paying peak-rate premiums.

Practical Ways to Reduce Gas Fees: Timing, Batching, and Transaction Settings

As of February 2026, a standard Ethereum transfer uses 21,000 gas, so shifting from 25 gwei to 15 gwei cuts the fee by 210,000 gwei (0.00021 ETH) per transfer. At an ETH price of £2,000, that change equals about £0.42 saved each time. Timing drives much of that spread because base fees track block-by-block demand, so sending during quieter periods reduces the market price for block space without changing what the transaction does.

Batching reduces fees by sharing fixed overhead across multiple actions. For example, combining two ERC-20 approvals and one swap into a single routed transaction can replace three separate base-fee payments with one, even if the total gas used rises. When setting fees, use a wallet’s “max fee” and “priority fee” controls to avoid overbidding; a 1 gwei tip on 100,000 gas costs 100,000 gwei (0.0001 ETH), while 3 gwei triples that amount. Cross-check current ranges on Etherscan Gas Tracker before signing.

Lower-Fee Alternatives: Layer 2 Networks, Sidechains, and Bridged Transfers

At 18:05 on a Friday, a user tries to bridge £200 of USDC from Ethereum to Arbitrum and sees a £6.80 Ethereum gas estimate for the deposit. The same user then swaps on Arbitrum for about £0.08 in fees, because the Layer 2 network batches hundreds of transactions and posts a compressed proof back to Ethereum.

That difference comes from where computation happens. Layer 2 networks (such as rollups) execute most activity off the Ethereum mainnet, then settle results on-chain, so users share the cost of one settlement transaction. Sidechains reduce fees by using a separate validator set, which changes the security model and can raise bridge risk. Bridged transfers also add a one-off cost: moving assets onto a scaling network often consumes 120,000–250,000 gas, depending on the bridge and token.

For routine transfers and trading, keeping assets on a reputable Layer 2 can cut per-transaction costs by 90% or more versus mainnet, while preserving Ethereum settlement. For context on why these trade-offs differ from Bitcoin’s design, see bitcoin vs ethereum explained.